If you use Amazon EC2 extensively, then you’ve fallen prey to the dilemma of “to reserve or not to reserve”. It is however a simple question in the sense that for a reserved instance, you pay a lesser per hour fee but end up paying an additional one-time fee upfront. So what is the optimal break-even point where you know you are not spending more than required? EC2 pricing page gives a high-level number for break-even point of Linux Reserved Instances.
The following table illustrates the percentage utilization needed for the break even point.
|
Utilization
|
1 year
|
3 years
|
|
Light
|
28%
|
11%
|
|
Medium
|
41%
|
19%
|
|
Heavy
|
56%
|
35%
|
Having gone through this hair-pulling situation and from our experience of using EC2 for a while, PromptCloud shares a guide that will help you better understand the AWS schemes that suit you best. You simply drag the “number of days/months” and “number of hours” sliders and it displays exact numbers on pricing. What better can you ask for an approximate use case :).
 |
| Reserved instance for 1 year |
 |
| Reserved instance for 3 years |
To join the discussion on Hacker News, go to here-https://news.ycombinator.com/item?id=5785911
2 replies on “Introducing Amazon EC2 On-demand vs. Reserved Instance Price Calculator”
Chris Adams
Great tool – one thought, though: rather than providing the separate days slider, what about simply displaying the full chart and displaying the axis labels so you can easily see where each of the reservation options becomes more expensive?
Eoin McMahon
This tool is invaluable. Instead of having to use the not so simple AWS Simple Monthly Calculator provided by AWS we can simply choose the server instance types that are most suited to our requirements in terms of CPU and Memory and quickly compare the costs of each instance type to decide on the best option. It would be cool if the tool could allow you to select a number of instance types across all the families to compare the on-demand and reserved costs of each.